Hi
I'm not really sure that I understood you correctly , but e.g. in ft peer configuration you have such options :
heartbeat interval X
heartbeat count Y
It means that device sends heartbeats with interval Xms and if Y heartbeats are missing from peer device, it will assume that it's down and take over. So, it should take around X*Yms to declare second device dead.
If you have query interface configured it will increase failover time, as device will try to check if peer is alive via this additional interface (
http://www.cisco.com/en/US/docs/interfaces_modules/services_modules/ace/vA5_1_0/command/reference/ft_peer.html#wp1031591 )
Then switchover happens pretty fast, transit connections are replicated. See limitations section :
http://www.cisco.com/en/US/docs/interfaces_modules/services_modules/ace/vA2_3_0/configuration/administration/guide/redundcy.html#wp1003562
